The GRAMMY® awards are on March 14th, and there’s a lot to learn by analyzing some of the records that have been nominated. But where to start? There are so many questions surrounding music production, mixing, and mastering, especially now that streaming services have started to move towards loudness-normalized playback. How loud should I make my master? What tonal balance should I aim for? How high should I set my limiter ceiling? How dynamic should my record be?
The answer is, of course, “it depends.” But in this article, we’ll see what the experts are doing so we can arrive at some answers for ourselves. To do so, we’ll take a technical look at the nominees for Record of the Year, which doesn’t only go to the artist, but also to the producer, mix engineer, and mastering engineer.
The first step was to buy uncompressed, lossless files for each of the nominees (from the Quobuz store). Next, we dropped those files into RX 8 and looked at the Waveform Statistics window. The sample peak shows the maximum sample level, which is where the mastering engineer set the ceiling of the limiter (in Ozone 9, this is the Maximizer module). True peak is the level of the analog signal after being converted from digital, which can be higher than the digital signal. If the mastering engineer used a “True Peak” limiter, then the sample peak should be equal to the true peak.
Integrated loudness is a measurement of loudness for the entire audio file. This is what streaming services will typically use if they’re normalizing tracks. Most streaming services normalize to -14 LUFS, so if a song is -8 integrated LUFS it will be gained down 6 dB. In addition to the integrated loudness, we’ll look at the loudness range, which is a measure of how dynamic a track is.
To analyze the tonal balance of the nominees—the distribution of signal across the frequency spectrum—we used Tonal Balance Control 2. To learn about how and when to use Tonal Balance Control, head here.
I took 15-sec sections from each of the nominees, during the fullest parts of each track (typically the final chorus), and used those to make a new target curve. We’ll start by taking a look at this target curve made with all the nominated tracks, and we’ll then look at each individual track’s tonal balance (represented by a white line) relative to that target curve.
Above, you can see a GIF fading between the “Bass Heavy” target curve (a factory target curve that comes with Tonal Balance Control) and the curve made from the Record of the Year nominees. There are a couple of interesting differences:
The GRAMMY® curve has more bass in general than the Bass Heavy curve (even overlapping the Crest Factor meter!)
Low-mids are pretty consistent except for a bump around 300 Hz
High-mids on the GRAMMY® curve definitely have less energy
The target area for the highs is much wider, indicating a wide variance in the GRAMMY® nominees
Now that we’ve taken a look at the general tonal balance of this year’s nominations, let’s take a closer look at each nominee, diving into both the tonal balance profile and waveform for each.
Producer: Beyoncé & Derek Dixie | Mixing: Stuart White | Mastering: Colin Leonard
True Peak: +1.01 dB | Sample Peak: -0.01 dB | Int. Loudness: -8.5 LUFS | Loudness Range: 2.7 LU
The first thing you notice about “BLACK PARADE” is just how big the waveform is (take a look at some of the other nominees’ waveforms below). Sample peaks go almost all the way up to zero, with true peaks of just above +1 dB. The track is consistently loud across the whole song, and its loudness range of 2.7 LU is the lowest of all the nominees. At the same time, “BLACK PARADE” is far from the loudest track at -8.5 LUFS (it’s the sixth loudest). It’s impressive how such a consistently loud track with full level achieves a relatively low integrated LUFS, meaning streaming services won’t turn it down excessively during any normalization.
The tonal balance is distinct for how low the bass extends. I checked, and that’s a B at 30.8 Hz. The other notable bump around 4kHz comes from Beyoncé’s vocal and a very present rolling hi-hat.
Producer: Adrian Quesada | Mixing: Adrian Quesada | Mastering: JJ Golden
True Peak: -0.38 dB | Sample Peak: -0.49 dB | Int. Loudness: -9.9 LUFS | Loudness Range: 10.3 LU
Right away, we can see that this is a very different-looking record. “Colors” is the quietest record among the nominees at -9.9 LUFS, and is also very dynamic with 10.3 LU of range. This means that “Colors” will be turned down less than all the other records in this list if loudness-normalized. We can also tell that the mastering ceiling was set at -0.49, leaving some headroom for true peaks and level changes from lossy encoding.
The waveform has drum transients punching out, not a thick, sustained 808 like in “BLACK PARADE.” The resulting tonal balance is light on bass and high frequencies, but has a very present midrange occupied by the vocals, bass guitar, keys, and acoustic drum kit.
Producer: SethInTheKitchen | Mixing: Derek "MixedByAli" Ali, Chris Dennis, Liz Robson, Chris West | Mastering: Glenn A Tabor III
True Peak: -0.43 dB | Sample Peak: -0.47 dB | Int. Loudness: -8.4 LUFS | Loudness Range: 10.1 LU
The waveform for “Rockstar” is distinct for its thick 808 sections contrasted against the more quiet verses and breaks, creating a very dynamic track with a loudness range of 10.1 LU. It is also in the lower half of integrated loudness, meaning those 808 sections will really hit when normalized. The peak levels are also set fairly low. Tonal balance shows a big emphasis on bass dominated by the 808 kick. There’s also a bump around 8kHz occupied by the rolling hi-hat and the high-end fizz of the vocals.
Producer: Tyson Trax | Mixing: Clint Gibbs | Mastering: Mike Bozzi
True Peak: +0.00 dB | Sample Peak: -0.09 dB | Int. Loudness: -7.0 LUFS | Loudness Range: 5.5 LU
“Say So” by Doja Cat is a consistently loud track, with the exception of the intro and outro which are both low-pass filtered. The peak levels are set so that there is a perfect 0 dB true peak without using a true peak limiter. Tonal Balance Control shows a boosted high end coming almost entirely from Doja Cat’s super bright and shiny vocals.
Producer: Finneas O'Connell | Mixing: Rob Kinelski, Finneas O'Connell | Mastering: John Greenham
True Peak: -0.55 dB | Sample Peak: -0.60 dB | Int. Loudness: -9.4 LUFS | Loudness Range: 12.0 LU
Billie Eilish and her brother/producer FINNEAS return to the Record of the Year category after winning last year (along with four other categories.) They’re also working with the same mixing and mastering engineers. This record is an outlier among the nominees in a couple of ways. “everything i wanted” is the most dynamic record with 12 LU of range. The sample peak is also set the lowest of any nominee. The tonal balance has the most bass and the least highs and high mids, creating a tonally dark mix that matches the dark vibe of the song.
Producer: Caroline Ailin, Ian Kirkpatrick | Mixing: Josh Gudwin, Drew Jurecka, Ian Kirkpatrick | Mastering: Chris Gehringer
True Peak: +0.27 dB | Sample Peak: -0.10 dB | Int. Loudness: -7.9 LUFS | Loudness Range: 5.2 LU
The waveform for Dua Lipa’s “Don’t Start Now” shows a consistently loud song with some clear dynamic differences between sections with and without the kick drum. Sample peaks are set to -0.10 dB, which lets some true peaks come through. The tonal balance has elevated high-mids and highs which are packed with percussion and Dua Lipa’s vocals.
Producer: Louis Bell, Frank Dukes, Post Malone | Mixing: Louis Bell, Manny Marroquin | Mastering: Mike Bozzi
True Peak: +0.14 dB | Sample Peak: -0.04 dB | Int. Loudness: -5.6 LUFS | Loudness Range: 5.4 LU
Post Malone and his production team are returning to the Record of the Year category after last year’s nomination for “Sunflower.” “Circles” is the loudest of the Record of the Year nominees at -5.6 LUFS. I feel like the more aggressive loudness fits the vibe of the track which is a bit more lo-fi rock, featuring a crushed and saturated drum kit. In terms of tonal balance, this track has a higher bass around 150 Hz, rather than a booming and low 808. It also has present low-mids which are mainly Post Malone’s vocals, and a peak around 10 kHz which is the high-end snap of the drums.
Producer: Beyoncé, J. White Did It | Mixing: Stuart White | Mastering: Colin Leonard
True Peak: +0.80 dB | Sample Peak: +0.00 dB | Int. Loudness: -8.0 LUFS | Loudness Range: 4.5 LU
We have another song with Beyoncé, remixing Megan Thee Stallion’s “Savage”. A couple fun GRAMMY® facts: this is the second time an artist has occupied two nominations for Record of the Year in the same year. The first was Pharell for “Get Lucky” and “Blurred Lines” in 2014. Also, Beyoncé now has seven Record of the Year nominations, tied for the most ever with Frank Sinatra. She really deserves the title Queen B!
This record is very consistently loud with 4.5 LU of loudness range, but you can still see very pronounced kick transients in the waveform. It is the only track with a sample peak all the way up to 0.00 dB. The high end is particularly pronounced because of an open hi-hat sample in the loudest part of the song.
So now that we’ve seen all of this data, what have we learned? The first trend that I notice is that the max true peaks are all higher than the max sample peaks. In fact, half of the nominees had true peaks above 0 dB. Based on the sample peaks, it doesn’t look like they’re worried about the clipping that can happen with lossy codecs.
There's a fairly consistent range in terms of integrated loudness, from Black Pumas at -9.9 LUFS to Post Malone at -5.6. The average integrated loudness for the nominees is -8.0 LUFS. Now that some streaming services have normalization options, you can make the decision to have a more punchy record like “Colors” or a more consistent wall of sound like “Circles”. Loudness range within songs varied between the consistently loud “BLACK PARADE” and highly dynamic “everything i wanted”. This definitely seems like a creative choice that should fit the feel of the song.
Tonal balance varied significantly between nominees this year. We have bright, shiny pop songs like “Say So” by Doja Cat, where the vocals have a really prominent high end. On the other hand, “everything i wanted” has a much darker, bassier balance to match the record’s brooding tone. So the question of tonal balance is really one where the answer will depend on your track or genre. Fortunately, the genre presets in Tonal Balance Control 2 can tell you just how you measure in comparison. You can also create your own reference curves from your own selection of tracks, like I’ve done here.
Looking forward to another analysis in 2022!